The most recent Legends & Lore column (this one is by Mike Mearls), is about save or die, one of those perennial fault lines within the D&D community. A few notable blogs have already commented on it (Discourse & Dragons, Tenkar’s Tavern, The Nine and Thirty Kingdoms). The presence of save or die traps is also one of my 20 rules clarification questions because it is an important aspect of how a particular game plays (as I expected, based on my readership, most of the people that posted their answers to those questions responded that save or die was an active threat in their game).
Why is this such a divisive issue? One, players don’t like their characters to die. That is clearly not the entirety of the problem though. A dragon’s claws, or a enemy’s sword, can also end the life of a PC (instantly, in most versions of D&D, if enough damage is done). Two, the save or die mechanic emphasizes the fact that only one line separates the PC from death. HP, in many cases, might be the same thing in practice (though not in all editions), but it doesn’t feel the same, because there are more rolls involved (at least the attack roll and damage roll) and the defense (AC) is somewhat under the control of the player, whereas saves are either entirely based on level (traditionally) or mostly based on level. Back in my Second Edition days, we didn’t use save or die (or level drain). As Robert Fisher notes, this was probably due to my lack of a Gygax Number (that is, the lack of a mentor connected to the original play style). Instant kill effects felt silly and unfair to us, but that was because we didn’t understand how they were supposed to be used.
My position now is that rolling a saving throw means you have already done something wrong. The save is a second chance, and the alternative is to just die. (I have been most heavily influenced here by -C from Hack & Slash.) However, this has an important corollary that is usually not discussed: there is such a thing as an unfair save or die effect. Just because the player has a stat on a character sheet and is rolling a die based on some rules does not mean that the situation is fair. There are certain (mostly unwritten) covenants that guide the usage of this most potent of mechanics.
Every fair save or die effect should be avoidable by smart play. If it is not, the effect is unfair. For example, PCs should know (or be able to learn) that a particular monster has a gaze attack. This could be just assumed (“you have heard of the fearsome basilisk gaze”) or it could be something that the party can ask the one-armed man in the tavern about. Traps leave marks on their environment. Scorched sections of wall. The heaped bones of past victims. Abnormally clean sections of flooring. Given information and planning, PCs can also prepare by preparing countermeasures, such as buying antidotes (something I would not make too difficult, as it reflects players engaging with the threats in the campaign world).
Sometimes context is enough of a clue. Extreme example: exploring a dwarven machinist’s trap factory. Why would a sane person touch anything without precautions in such an environment? Most games should not consist entirely of such environments though, for the same reason that most settings should not feature only one kind of monster. It’s boring. And that, of course, is the worst sin that any game designer can commit.
This sounds good in theory, but there are a number of practical objections, best illustrated by unearthing some comments from a very interesting three part discussion of save or die effects over at Grognardia from back in 2009 (part 1, part 2, part 3). First, what about wandering monsters? Vedron writes:
However, I find that all too often in actual play save or die gets perverted. For example, the PCs might enter a combat with a randomly encountered medusa that can turn them to stone if they fail a save.
Ideally, there would be some warning of such an encounter; either there would be a rumor in the tavern, or there’d be some odd statuary, or some sort of reasonable recon could have ascertained the nature of the possible threat. If the players choose to proceed without protection or intelligence collection, so be it.
But, all too often, the GM just throws one in randomly as a surprise with no forewarning and rubs his hands in glee as PCs get stoned (or level drained with a wight; or slimed by green slime falling from nowhere; or crushed by instant death cave ins from nowhere, etc).
In this case, I would say that the PCs are adventuring an a locale that is too dangerous for them. Thus, they have already made a mistake. Why might this be unfair? If the referee has not provided other locations that can be explored, or has not provided any way that the players can learn about the threat level, or does not make it possible to flee from the encounter before engaging, then this is an unfair situation.
What about monsters that present a disproportionate threat? There is some idea floating around that only “big” (boss?) monsters should be able to use such devastating attacks. This can be shown in Mike Mearls’ Tiamat example in the Legends & Lore column linked above. For another example, Brunomac writes:
Not speaking of other affects, but I always thought of dying if missing a poison save was always a bit of a rip-off. So I usually base damage on the hit dice of the creature involved (although giving small very poisonous bugs bonus’), and let the damage pan out over several rounds.
I agree with Mearls and Brunomac in the following way: the reason Tiamat having a save or die effect seems “more fair” comes back to the idea that players should be able to learn about the threats they face. However, the play requirement of engagement with the setting to acquire that information has been circumvented in the case of threats like Tiamat, because, you know, TIAMAT. Players are already scared and on guard. However, I think the fix here is to make giant spiders more terrifying, not less dangerous. Think about Shelob from The Lord of the Rings. Think about how you might react to a sofa-sized arachnid with pincers dripping poison.
There are other techniques that can be used to communicate threat information to players. For example, you can make liberal use of red shirts. Kill retainers first. Then, if PCs still charge forward, guns blazing, they can’t say they weren’t warned. This might seem like going easy on players, and to some degree it is, but only because you are requiring less engagement and creativity for successful information gathering. The actual threat has not been reduced, and I think that is the most important part.
All that being said, save or die is somewhat incompatible with some new styles of play. To some players, D&D is about finding things to fight. This is not a problem with the mechanic though, it is a problem with player expectation about the nature of the game they are playing. I’m not going to say that people who want only a combat game out of D&D are wrong, but I will say that they are not playing the game to its full potential.